Through the bench test, the vibration characteristics of the front right support of the inline four-cylinder diesel engine is analyzed, and the MATLAB simulations and verifications are performed to the differential equation of 1 / 4 vibration model established by using the front right support as the research object. Experimental and simulation results show that the exciting force performs as basic frequency and second frequency and, the acceleration performs as basic frequency and low-frequency at low speed engine vibration; when it is in the high speed vibration, the exciting force is mainly manifested in the second frequency and the acceleration in high-frequency. The motion equation of 1/4 vibration model can reflect the status of the engine vibration well in the second frequency.

Engine fault diagnosis and detection is inseparable from the analysis and examination of each sensor or actuator, and the waveform display is the most intuitive and convenient way. This system is an engine waveform tester developed based on virtual instrument and the batch estimate fusion theory is used in the process of the single sensor data acquisition and processing. It is proved by practice that this instrument can conveniently and quickly realize the functions such as signal acquisition and control, waveform analysis and processing and result expressing and output, thus providing technical support for comprehensive intelligent engine fault diagnosis technology.

The fault data of 133 Xiali taxis in Xingtai Car Rental Company are collected and the preliminary sorting is conducted to them. The automobile preventive repair model is established and the analysis of the necessity and feasibility of preventive repair is performed. In the base of reliability analysis of fault data, the failure rule is obtained and the optimal repair cycle is found out.

No abstract

The nonlinear and hysteresis characteristics showed by magneto-rheological (MR) mount make it seem very difficult to establish a precise mathematical model. Based on the testing of MR mount dynamics, RBF neural network model can train and forecast the collected data. Analysis of comparing the predicting result of the RBF neural network model with the testing result shows that the trained RBF neural network model can exactly predict the dynamics of MR mount, and it provides some new ideas to implement the better intelligent control of the engine MR mount.